Senior Systems Software Engineer, Accelerated Kubernetes Performance and Scale - DGX Cloud

NVIDIA has been transforming computer graphics, PC gaming, and accelerated computing for more than 25 years, driven by great technology and amazing people.

We’re now tapping into the unlimited potential of AI to define the next era of computing, where our GPUs power computers, robots, and self‑driving cars that can understand the world.

Doing what’s never been done before takes vision, innovation, and the world’s best talent.

As an NVIDIAN, you’ll work in a diverse, supportive environment where people are encouraged to do their best work and grow their careers.

We offer a preference for hybrid work while remaining open to remote arrangements, giving you flexibility in how you do your best work.

Come join the team and see how you can make a lasting impact on the world.

The DGX Cloud organization at NVIDIA brings together cutting‑edge hardware and software innovation to deliver industry‑leading accelerated computing for the world’s most ambitious AI workloads.

We are a group of forward‑thinking engineers tackling some of the globe’s toughest challenges, pushing progress, and positively affecting millions of lives.

We’re searching for a Senior Systems Software Engineer with deep expertise in distributed systems, Kubernetes, containers, and systems performance and scalability.

The ideal candidate brings broad, hands‑on experience across the stack, including GPU operators, device plugins, distributed inference serving, and major cloud platforms.

You’ll own hard technical problems at large scale and help shape how AI infrastructure runs in production.

In this key role, you will focus on scaling AI infrastructure while minimizing total cost of ownership, reducing cost per token and enabling future AI innovation and AI factories.

Are you ready to be impactful?

What you'll be doing:

• Lead end‑to‑end performance and scalability analysis across the Kubernetes‑based accelerated runtime stack (control and data planes), including NVIDIA components such as GPU Operator, Network Operator, node-feature-discovery, topograph, dra-driver-nvidia-gpu, and nvsentinel, tracking issues from orchestration down to the metal.
• Design and contribute upstream architectural changes to the Kubernetes control plane and related projects to enable reliable operation at hyperscale cluster sizes, doing in the open what today’s hyperscalers typically do privately.
• Improve container startup and cold‑start latency to enable smooth, low‑latency inference scaling on Kubernetes across thousands of GPU nodes, ensuring the AI runtime stack scales without creating API server pressure or operational fragility.